Point to the exact claim.
Email data@chinaautoledger.com with the page URL, field or value, expected correction, official supporting source and access date. We do not accept an unsupported replacement value.
Reproduce before changing.
We compare the report with the retained source snapshot and mapping contract. A valid correction receives a new event and verification timestamp; the prior captured fact remains traceable rather than silently overwritten.
Disagreement is not equivalence.
Different official datasets may count different populations or periods. When both values are accurate within their own definitions, we preserve both and explain why they are not directly comparable.
Every report receives one evidence outcome.
A report may be confirmed, rejected, clarified or held for more evidence. A confirmed error changes the public current record and appends a dated correction trail. A rejected report leaves the record unchanged and documents why the submitted evidence did not alter the published claim.
A clarification improves the surrounding definition or explanation without pretending the underlying official fact changed. A held report remains unpublished when access rights, mapping or source continuity cannot yet be verified. The website never substitutes an estimate while a review is open.
Correction requests should not ask us to merge unlike measures. Registered stock, first registrations, sales, deliveries and market share describe different subjects. The correction process protects those distinctions as carefully as it protects the original number.